# Fix list: Capsule CRM From Anchor Terminal's listing at https://www.anchorterminal.com/tools/capsule-crm, the October 2026 research run, assessed 8 October 2026. Grade C, 55.7 out of 100. This is everything the published grade says the listing lacks, the biggest possible gain to the total first. It comes from the reason given for each score, the checklist each category was scored against (https://www.anchorterminal.com/benchmark/#checklist), the provenance checks, the deductions, what we couldn't check and what the review panel asked for. A fix counts at the next check, once it's public. For a coding agent working on Capsule CRM: work through the items below in the product, its docs and its public pages. Each category gives the reason for its score, with the points each checklist item earned, and the checklist itself, so the gap is the items that earned less than their points. Change the product, not the wording, and keep a note of what you changed and where it's published. ## 1. Security & auth, 46 out of 100, up to 9.5 more on the total Why it scored 46: OAuth 2 authorisation code grant with scopes `read`, `write` and `user_preference`, access tokens of about seven days, refresh tokens, a revocation endpoint and S256 PKCE in the server metadata. Personal tokens are also accepted on the REST API, and their scope is not documented. Tokens travel only in the header (24 of 30). A `read` scope exists, and the MCP server acts with the connecting user's access controls. No server-side confirmation for writes was found, and the AI Terms put review of each tool call on the customer (12 of 20). Notes and stored emails reach the model, and no injection guidance was found (2 of 15). No API audit log was found, and the support page says MCP update tools keep no change history of their own (2 of 15). A reporting address (support+sensitive@capsulecrm.com), regular penetration tests by outside specialists per the security page and PCI-DSS for card collection. No security.txt, no bug bounty and no certification named on the pages read. The trust centre at trust.zestia.com is drawn by script and was not read (6 of 20). The checklist (https://www.anchorterminal.com/benchmark/#checklist-security): - 0 to 30, the credential model. 30 for OAuth 2.1 with scopes, or scoped and revocable keys with rotation. 20 for plain revocable API keys. 10 for one all-powerful key. 10 off when a secret can travel in a URL query string as a documented option. - 0 to 20, read-only or least-privilege modes, and confirmation or approval for destructive actions. - 0 to 15, prompt-injection posture where the tool returns untrusted content (documented mitigations or guidance). A tool that returns no untrusted content gets 10. - 0 to 15, audit logs or per-call visibility for the operator. - 0 to 20, a security programme. security.txt or a disclosure policy, a bug bounty, SOC 2 or ISO 27001, advisories handled in public. Models are read for retention, whether API data trains models (and whether that's off by default), zero-retention options and certifications. Frameworks for telemetry defaults, approval hooks, guardrails and sandboxing. ## 2. Payments & pricing, 30 out of 100, up to 8.8 more on the total Why it scored 30: No x402, MPP or L402 (0). Plans are public but priced per user, with nothing per call, and the pricing page draws the numbers by script. The Starter price of $18 and the $0 to $72 range came from Capsule's comparison pages (10). A Free plan for two users and 250 contacts and a 14-day trial with no card. Whether the Free plan includes the API was not found (20). A person signs up in a browser and creates the token, and registering an OAuth app needs a GitHub sign-in (0). The checklist (https://www.anchorterminal.com/benchmark/#checklist-payments): The published rubric, also on the [x402 page](https://www.anchorterminal.com/x402/). - 40, a machine payment protocol (x402, MPP or L402) on the tool's own endpoints. 10 to 30 when it covers only some endpoints or only goes through a third party, and the note says which. - 20, per-call or per-unit pricing published without a login. 10 for public plan-only pricing, 0 for "contact sales" or prices behind a login. - 20, a free tier or trial that doesn't need a card. - 20, autonomous onboarding, meaning an agent can get access without a person signing up in a browser (keyless use, x402, a programmatic key API). Payment platforms and agent wallets rarely charge for their own API over a machine protocol, so the first line has steps for them, and the highest one that applies counts. 40 when x402, MPP or L402 runs on all their own endpoints, 30 when it runs on part of their own API, 25 when their merchants can accept one, 20 for running a facilitator, 15 for paying as a buyer, and 0 when the only protocol is their own. Merchant acceptance sits above a facilitator because the platform's own customers can charge agents through it, while a facilitator settles for sellers who wire up the protocol themselves. The counter-argument (a facilitator does more for the protocol as a whole) has a point. Each note says which step applied. Open-source software you run yourself is scored on its hosted or paid option if it has one. A free, self-hosted package with nothing to buy gets 20, 20 and 20 for the last three lines, and 0 to 40 for the first only if it ships a payment protocol. ## 3. Schema & documentation, 48 out of 100, up to 8.5 more on the total Why it scored 48: No OpenAPI or other machine-readable contract was found. Four likely paths on developer.capsulecrm.com returned 404 and the docs link none. The MCP server's tool schemas sit behind OAuth and were not read (0 of 25). No llms.txt on either domain (0). Each of the 132 operations has a short description, and some say when to prefer another endpoint, such as filters over search (12 of 20). Parameter tables give types, ranges and defaults, and each resource has a model page, with filter conditions as JSON bodies (10 of 15). Request and response examples on every operation, and an errors page with bodies for 400, 401, 403, 422 and 429 (13 of 15). A `/api/v2` path and a dated changelog with five entries in the last 12 months (13 of 15). The checklist (https://www.anchorterminal.com/benchmark/#checklist-schema): APIs and MCP servers. - 25, a machine-readable contract (a public OpenAPI file or similar; for MCP, typed JSON Schema inputs on every tool). - 10, llms.txt or Markdown docs served for agents. - 0 to 20, descriptions that say what a tool is for, when to use it and when not to, read from the tool definitions in the source or the API reference. - 0 to 15, typed inputs with enums, constraints and required fields, and no free-form JSON blobs. - 0 to 15, examples and documented error responses. - 15, versioning and a public changelog. Models are read from the API reference, the OpenAPI file, llms.txt, the structured-output and tool-use docs and the model cards. Frameworks from docs a model can follow, typed interfaces, examples and the API reference. ## 4. Agent ergonomics, 59 out of 100, up to 6.7 more on the total Why it scored 59: Lists return a default record and take `embed` to add tags, custom fields and related data, with `perPage` from 1 to 100. There is no field selection. The MCP reference lists 37 tools, which is above the 30-tool line (15 of 25). `page`, `perPage` and a `Link` header, `since` on lists, `/search` endpoints, structured filters and lists of deleted records (18 of 20). Errors are JSON, a 422 names the resource and field, and 403 states the missing scope (16 of 20). No idempotency keys. Updates are partial PUTs, which are safe to repeat, and MCP tool annotations were not readable (4 of 20). Few required fields and sensible defaults, but no official SDK in any language (6 of 15). The checklist (https://www.anchorterminal.com/benchmark/#checklist-ergonomics): - 0 to 25, context cost. For MCP, the number and size of the tool definitions (25 for ten or fewer compact tools, 15 for 11 to 30, 5 for more than 30, plus up to 10 back for toolsets, dynamic loading or read-only subsets). For APIs, whether responses can be sized (field selection, limits, summaries). - 20, pagination, filtering and output-size controls. - 20, actionable, documented error responses, codes and messages an agent can recover from. - 20, idempotency or safe retries, and for MCP the `readOnlyHint` and `destructiveHint` annotations. - 15, sensible defaults, few required parameters, and official SDKs in at least two languages. Models are read for tool use, structured output, prompt caching, context length, batch and SDKs. Frameworks for how much code and how many defaults a tool-calling agent with MCP needs. ## 5. Maintenance & community, 38 out of 100, up to 5.4 more on the total Why it scored 38: Last change on 25 September 2026, the hosted MCP server, 13 days before the check (30). Two dated entries in the last 90 days, 19 August and 25 September 2026, which is short of three (0). A public changelog, a support address and a Google group for API questions. We did not test support (8 of 15). No official SDK, and the MCP server is not in the official MCP registry (0). No packages to judge (0). The checklist (https://www.anchorterminal.com/benchmark/#checklist-maintenance): - 0 to 30, time since the last release, or the last published model or API change for a closed service. 30 within 30 days, 20 within 90, 10 within 180, 0 older. - 20, at least three releases or dated changelog entries in the last 90 days. - 0 to 25, responsiveness. Issues and pull requests answered on GitHub (the open issues and how recent the replies are). For closed services, a public changelog and a support or community channel that answers, 0 to 15. - 15, presence in the official MCP registry under a verified namespace (MCP servers), or current official SDKs (APIs and models). - 10, package health, current dependencies and CI. Models are read for deprecation notice periods and model churn rather than release counts. ## 6. Reliability, 83 out of 100, up to 3.4 more on the total Why it scored 83: Graded on REST API v2, with the hosted lines. Status page at status.capsulecrm.com on Statuspage with two components and history back to 2016 (20). No incident between 10 July and 8 October 2026. The last, a minor delay to reports and dashboards, ran from 6 to 8 July 2026, and a major outage of about 15 hours fell on 27 March 2026, both outside the 90 days (30). Limit published, 4,000 requests an hour per user (15). A 429 carries `X-RateLimit-Reset` and the docs say to wait until that time and suggest one second between calls, but there is no Retry-After header, no idempotency key and no safe-retry advice for writes (10 of 15). No SLA found, and the Customer Terms disclaim liability for unavailability (0). The REST API is generally available. The MCP server was announced on 25 September 2026 and the Claude connector is labelled beta on one support page (8 of 10). The checklist (https://www.anchorterminal.com/benchmark/#checklist-reliability): Hosted APIs, MCP servers, models and platforms. - 20, a public status page with component history (Statuspage, Instatus, BetterStack or the vendor's own). - 0 to 30, the incident record for the last 90 days on that page. 30 for a clean record or trivial incidents only, 20 for minor incidents only, 10 for one major outage (an hour or more of a core API down, or errors across the board), 0 for several. 5 when there's no history we could read, and the note says so. - 15, rate limits documented with numbers. - 15, documented 429 or overload handling (Retry-After, backoff guidance), and idempotency keys or safe-retry guidance where writes are involved. - 10, an SLA published for any paid tier. - 10, the surface agents use is generally available, not beta or preview. Local packages, SDKs, frameworks and stdio MCP servers. - 20, installs from an official package with supported runtimes stated. - 25, a public CI and test suite, passing on the default branch. - 0 to 25, open crash or regression issues relative to activity (25 for few and handled, 0 for many, old and unanswered). - 15, semver discipline and breaking changes called out in a changelog. - 15, version 1.0 or later, or declared stable. Protocols are read from their reference implementations, the public facilitators or servers, spec stability and test vectors. ## 7. Transparency & trust, 75 out of 100, up to 2.2 more on the total Made of editorial 64, provenance 86. Why it scored 75: Closed service with Customer Terms naming Zestia Limited, registered in England (15). The privacy policy of 21 August 2025, the DPA and the terms agree on a 14-day restore window after cancellation, deletion after that and backups for up to 50 days, with hosting on AWS in the United States. The AI Terms say inputs are not used to train models. The Claude connector page calls Anthropic a sub-processor, while the sub-processor list names OpenAI and not Anthropic (24 of 30). No deprecation policy. The changelog says Capsule tries to avoid breaking changes without notice, and the 19 August 2026 entry on the `pictureURL` change gives no date for the switch (8 of 20). Eleven sub-processors listed with location, transfer mechanism and safeguards, and 30 days' notice of changes in the DPA (17 of 20). The checklist (https://www.anchorterminal.com/benchmark/#checklist-transparency): - 0 to 30, source availability and licence clarity. 30 for open source under an OSI licence, 15 for closed with clear terms, 0 for unclear terms. - 0 to 30, data handling and retention statements that agree with each other (privacy policy, DPA, retention periods, subprocessors). - 0 to 20, a deprecation policy or notices with dates. - 0 to 20, telemetry disclosed with an opt-out (local software), or subprocessors and data locations disclosed (hosted). The other half of Transparency and trust is the provenance score, computed from checked facts (below). The category score is the mean of the two. Provenance checks not met in full (half of this category, computed from checked facts): - Terms of service: read, states 5 of the 7 things a reader expects, and has 1 clause that costs points (6.3 of 10) - security.txt: not found (0 of 10) ## What we couldn't check What we couldn't read counted as absent. Publishing it on a page a plain HTTP fetch can read (not only in a browser) lets the next check count it. - unchecked: plan prices on https://capsulecrm.com/signup/, which the page draws by script. Growth and Advanced prices and the monthly-billing prices were not read - unchecked: certifications in the trust centre at https://trust.zestia.com/, which is drawn by script - unchecked: the MCP server's tool definitions, annotations and exact tool count, which need an OAuth sign-in on a Growth plan - Whether the Free plan includes API access was not found in the reviewed pages - Whether personal API tokens can be limited to read access or revoked one by one is not documented - Capsule's pages disagree on what the MCP server can write, and on whether the Claude connector is in the directory or in beta - The Customer Terms carry no date on the page - No official GitHub organisation or SDK was found. github.com/capsulecrm returned 404 ## Weaknesses - No OpenAPI file, llms.txt or official SDK was found - The MCP server needs the Growth plan or above, and clients that only support dynamic client registration can't connect - Capsule's own pages disagree on what the MCP tools can write, and the tool reference lists only custom-field updates - No idempotency keys, and no Retry-After header on 429 - The pricing page draws its numbers by script. Only the Starter price and the top of the range were readable ## What costs an agent a turn today The notes we give agents before they call it. Each one is a workaround an agent shouldn't need. - Send `Authorization: Bearer ` to `https://api.capsulecrm.com/api/v2`. Projects use the old name in paths, `/kases` - Follow the `Link` header for the next page. `perPage` goes up to 100 and the default is 50 - On 429, wait until the epoch time in `X-RateLimit-Reset`. The docs suggest one second between calls - Ask for the `read` scope when the task only reads. A write with a read-only token answers 403 with `insufficient_scope` - For MCP, register an OAuth app in the developer hub first (GitHub sign-in needed) and give the client its ID and secret. Personal API tokens are refused ## When it's done Send what changed and where it's published as a dispute (https://www.anchorterminal.com/builders/#disputes, or `POST https://www.anchorterminal.com/api/v1/contact` with `"kind": "dispute"`). Disputes are answered in public, and the listing is checked again by the same checklist. Paying for an audit or a listing claim changes nothing here.